Seleção de intervalo de tempo para ajustar modelos de regressão às curvas de progresso e eficiência do drone no controle da ferrugem em cafeeiro Cultivar Catuaí 144

Abstract

Brazil is accounting for approximately 44% of global production. However, its productive capacity depends on several factors, including disease control. Among these, coffee leaf rust (Hemileia vastatrix) has the greatest destructive potential. Due to potential losses, studying epidemics under different management conditions becomes indispensable for improving decision-making. Thus, one of the objectives was to determine the best time interval for fitting regression models to the progress curves of coffee leaf rust in crops of the cultivar Catuaí 144, subjected to different management strategies. The progress curves were analyzed and time intervals were selected to study the epidemics. The choice of the best interval was based on the analysis of epidemiological variables and the representativeness of the main characteristics of the epidemic. Subsequently, linear and non-linear models (linear, quadratic, cubic, monomolecular, logistic, Gompertz, and exponential) were fitted to all progress curves. The time interval from February 11 to June 11, 2019, i.e., from 75 to 195 days after the start of assessments on November 13, 2018, showed the greatest potential for describing the epidemic under different disease management conditions. For the control treatment, the exponential model had the highest R2 of 0.83, a low AIC (66.65), and the lowest MSE (63.59). The exponential and cubic models were the best models for describing the temporal dynamics of the disease under management conditions. The second experiment aimed to evaluate the efficacy of spray applications using drones (Unmanned Aerial Vehicles - UAVs) in controlling coffee leaf rust, Cercospora leaf spot, and Phoma leaf spot on the coffee cultivar Catuaí 144, in different thirds of the plant and at two depths on the plagiotropic branches. Water-sensitive papers were used to analyze spraying efficiency, and epidemiological variables were used to determine disease control efficacy. The setups using the XR11015VS nozzle, with added adjuvant and a spray volume of 30 L/ha, were responsible for achieving the highest efficacy, with a deposition of 17%, 164.44 droplets/cm2, coverage of 2.38%, foliar deposition of 4.97 L/ha, and VMD <100 but with an AR of 0.65. In contrast, the turbo atomizer achieved between 372.9-522.0 droplets/cm2, coverage of 26.2-44.7%, deposition of 145.0-363.0 L/ha, and VMD between 337.0-742.0 μm. For rust control, the turbo atomizer achieved the best results, with the lowest RAUDPC. Although the UAV with 30 liters and added adjuvant showed no significant difference from the turbo atomizer in controlling Phoma leaf spot in the upper third of the coffee plant and for Cercospora leaf spot on average across the plant thirds, despite having lower droplet coverage compared to the turbo atomizer under the conditions of this experiment.